HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 1, 233-242, January 4, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/1/233    most recent M106678200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Alvarez, J. Articles by Domínguez, P. Search for Related Content PubMed PubMed Citation Articles by Alvarez, J. Articles by Domínguez, P. Social Bookmarking                      What's this?

Characterization and Cloning of Two Isoforms of Heteroglobin, a Novel Heterodimeric Glycoprotein of the Secretoglobin-Uteroglobin Family Showing Tissue-specific and Sex Differential Expression^*

Javier Alvarez, Jorge Viñas^§, José M. Martín Alonso, Juan Pablo Albar^¶, Keith Ashman, and Pedro Domínguez^**

From the  Departamento de Bioquímica y Biología Molecular, Edificio Santiago Gascón, Universidad de Oviedo, 33006 Oviedo, Spain, ^¶ Centro Nacional de Biotecnología, Consejo Superior de Investigaciones Científicas, Campus de Cantoblanco, Universidad Autónoma, 28049 Madrid, Spain, and  Samuel Lunenfeld Institute, Toronto MSG 1X5, Canada

Received for publication, July 16, 2001, and in revised form, October 26, 2001

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Heteroglobin (HGB) is a 39-kDa heterodimeric protein detected under non-reducing conditions in harderian, parotid, and submaxillary glands and saliva of the Syrian hamster with antiserum raised against the carboxyl end deduced from the female harderian gland cDNA FHG22 (Domínguez, P. (1995) FEBS Lett. 376, 257-261). After reduction, only one 5.6-kDa polypeptide, named HGB.A, was immunodetected and identified by sequencing as the mature FHG22 product. Tissue-specific expression of HGB.A and HGB mimics that of FHG22 mRNA, with sex differences in submaxillary and harderian glands. Purification of HGB revealed it consists of HGB.A disulfide bonded to HGB.B, a 33.5-kDa N-glycosylated subunit that yields a 9-kDa core polypeptide after deglycosylation. Two highly homologous (96.2%) cDNA clones (HGB.B1 and HGB.B2) encoding 94 amino acid-long isoforms were identified by screening a female harderian gland library with an HGB.B probe. The corresponding mature polypeptides are 78 amino acids long with 12 differences, but 3 putative N-glycosylation sites are maintained. The expression of HGB.B mRNAs is parallel to that of HGB and HGB.A, but no HGB.B2 mRNA was detected in submaxillary glands. Homology studies indicate that HGB.A and HGB.B1/HGB.B2 belong to different subfamilies of the secretoglobin-uteroglobin family and form heterodimers as previously described.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The existence of a uteroglobin/Clara cell 10-kDa family of proteins including UGB/CC10 orthologs and paralogs (such as subunits of rat prostatein, cat Fel d 1 and mouse androgen-binding protein, and cDNAs like hamster FHG22) was previously suggested (1-4). New related proteins and cDNAs were described (5-9), and the family was formally established during a meeting in which a nomenclature committee (10) coined the generic name of secretoglobins (SCGBs).¹ This family includes a diverse group of small, -helical, secreted polypeptides (10-12) only described in mammals and reported to form dimeric structures bound by interchain disulfide bridges involving two or three conserved Cys residues (4, 8, 12). Five or six subfamilies have been defined by homology rather than by functional features (10, 13), in agreement with reported specific dimeric associations between members of subfamilies (13). Some SCGBs have also been shown to form heterotetrameric associations in which two disulfide-bound heterodimers are non-covalently bound (8, 14-16). A tissue-specific expression pattern linked to exocrine epithelia has been found for all the members (2-10) whose levels can also be regulated by hormones (4, 17-20), including sex steroids also found to be ligands for some SCGB oligomers (2, 12, 14, 15, 17, 21). The only homodimer and best-studied protein of the family is UGB (4, 17, 22, 23), for which several ligands have been described, including progesterone/steroids (24, 25), other hydrophobic ligands (12, 26-28), retinoids (29), and calcium (27, 30). Contrarily to calcium, it has been shown that lipophilic compounds bind to an internal cavity formed between the two polypeptides of the UGB homodimer (17, 27, 28), and the existence of such a hydrophobic pocket in heterodimeric SCGBs has also been proposed (10, 12). Several groups report UGB binding to cellular and matrix proteins and to a possible membrane receptor (31-33). Besides reports of different cellular and physiological actions (4, 17, 21), some of which arise from knock-out projects (34-36), the physiological role(s) of UGB and, in general, SCGBs is unclear.

We had previously prepared two sex-differential cDNA libraries and isolated male- and female-specific clones (3, 37) from Syrian hamster harderian glands. These are secretory organs from the orbital cavity related to the pineal gland and the gonads (38, 39), which in hamster show reversible sexual dimorphism regulated by hormones and other factors (40-42). The female harderian gland cDNA clone FHG22 was characterized in our laboratory and related to the UGB family as mentioned (3). The FHG22 mRNA was found to be expressed according to a tissue- and sex-specific pattern (3) only in three hamster exocrine glands; the highest levels are observed in parotid glands from either sex, and FHG 22 mRNA is present in female but not in male harderian glands and presents higher expression levels in female than in male submaxillary glands. These sexual differences led us to develop studies on hormone regulation; estradiol stimulates FHG22 mRNA expression in harderian glands both in vivo and in vitro, whereas no effect is observed using other sex steroids (20).

In this paper, we report the use of an antipeptide antiserum that specifically recognizes HGB.A, the product of the FHG22 mRNA, in monomeric or oligomeric form. HGB.A is the small subunit of a disulfide-bound heterodimer named heteroglobin (HGB), also formed by the large N-glycosylated subunit HGB.B. Sequencing of HGB.B enabled us to isolate two cDNAs corresponding to highly homologous isoforms (HGB.B1 and HGB.B2). HGB.A and HGB.B belong to different subfamilies of the SCGB family, and their mRNAs show a sex- and tissue-specific expression identical to that of HGB, but HGB.B2 is surprisingly absent in submaxillary glands.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Animals-- Male and female Syrian hamsters (Mesocricetus auratus) were obtained from Charles River (Kingston, NY) and maintained under controlled temperature (20 ± 2 °C) and photoperiodic conditions (14:10 h, light/dark cycle) with food and water ad libitum. The animals were sexually mature (about 4 months old) when used for experimentation. For preparation of tissue homogenates or RNA extraction, the animals were killed by suffocation with carbon dioxide, and the tissues were rapidly dissected and washed with ice-cold phosphate-buffered saline. For the preparation of antipeptide antisera, outbred New Zealand male rabbits were maintained in controlled conditions for 40 days before the immunization protocol was started and bled when needed.

Preparation of Hamster Saliva Samples and Tissue Homogenates-- Hamster saliva samples were obtained by inserting sterile cotton ear buds inside the animals' mouths and allowing them to chew for 2 min. The cotton plugs were removed, inserted into a bottom-cracked 0.5-ml tube placed inside a 1.5-ml tube, and centrifuged at 15,000 × g for 5 min. Clarified saliva was collected, and the plugs were washed with one volume of buffer A (0.15 M NaCl, 10 mM EDTA, 0.1 µM phenylmethylsulfonyl fluoride, and 1 µg/ml of aprotinin in 50 mM Tris-HCl pH 8.0) and centrifuged. Finally, each sample of saliva and eluted buffer A was mixed and centrifuged again for 10 min to precipitate the remaining cells and debris. The protein concentration of each saliva sample (average value, 6 mg/ml) was measured by the Bradford dye binding assay (Bio-Rad) (43) using bovine serum albumin (Sigma) as a standard. Cotton ear buds, bottom-cracked tubes, and buffer A were also used to collect samples of female hamster vaginal discharge.

To obtain the homogenates from parotid, submaxillary, and harderian glands, the corresponding organs from at least 2 animals were dissected as mentioned, mixed, and homogenized with a Polytron in 2 volumes of an ice-cold solution containing freshly prepared buffer A. The preparations were then centrifuged at 15,000 × g for 60 min at 4 °C, and each supernatant was collected and denominated parotid, submaxillary, or harderian homogenate. Protein concentration was measured as above (43), with average values of 19, 40, and 13 mg/ml for parotid, submaxillary, and harderian homogenates, respectively. These preparations were either used immediately or kept frozen at 70 °C until required. An identical preparation protocol was followed to obtain a thymus homogenate.

Preparation of Antipeptide Antisera-- The peptides CPAVLSVSKSFLFDKVEKFEC (FHG22-(24-55)), CEAVKAKVEVKKC (FHG22-(55-66)), and KMEMGKILAEVVGYCKGTEN (FHG22-(76-95)), corresponding to amino, central, and carboxyl sections of the FHG22 ORF (3), were synthesized with an AMS 422 automated multiple solid phase peptide synthesizer (Abimed, Langenfeld, Germany) using standard Fmoc (N-(9-fluorenyl)methoxycarbonyl) chemistry (44). The peptides FHG22-(24-45) and FHG22-(55-66) contain an extra Cys with respect to the deduced sequence in the carboxyl and amino terminus, respectively, for coupling purposes. The three peptides were coupled to keyhole limpet hemocyanin (Pierce) via Cys using the linking agent Sulfo-succinimidyl 4-(N-maleimido-methyl)cyclohexane-1-carboxylate. These conjugates were used as antigens to induce production of antibodies against polypeptide HGB.A (see "Results") by immunization of New Zealand rabbits. Animals were injected subcutaneously in multiple sites with a preparation containing 500 µg of the conjugate emulsified with an equal volume of Freund's complete adjuvant (Difco) in a total volume of 1 ml. Two intramuscular boosts of 200 µg of the conjugated peptide emulsified with incomplete adjuvant were given 4 and 7 weeks later, and sera were collected 2 weeks after the last injection. For the preparation of sera, blood was extracted from each rabbit, left to coagulate, and centrifuged for 30 min at 2,000 × g. The antipeptide antisera were finally mixed with 1 volume of 87% glycerol, separated into aliquots, and stored at 20 °C.

Protein Electrophoresis and Western Blot Analysis-- Denaturing protein electrophoresis (SDS-PAGE) was performed according to the method of Laemmli (45) with some modifications. For non-reducing SDS-PAGE, samples were mixed 3:1 with loading buffer containing 8% SDS, 8 mM EDTA, 40% glycerol, and 0.01% bromphenol blue in 0.25 M Tris-HCl, pH 7.5. For reducing or semi-reducing SDS-PAGE, samples were mixed 3:1 with loading buffer and treated, respectively, with 1.4 or 0.14 M -mercaptoethanol at room temperature for 30 min. Samples were finally heated at 100 °C for 10 min, electrophoresed in 10% or 15% analytical SDS-polyacrylamide gels, and either stained with Coomassie Brilliant Blue R-250 or transferred to nitrocellulose. The apparent molecular weight of protein bands was measured by interpolation from a linear semilogarithmic plot of M_r versus distance of migration using the MultiMark^TM multi-colored mixture of protein standards (Novex, San Diego, CA). For Western blot analysis, proteins were electrophoretically transferred from the gels to Duralose UV^TM (Stratagene) essentially according to Towbin et al. (46), and membranes were stained with 0.01% Ponceau S in 0.1% acetic acid to verify protein transference. Blots were blocked in phosphate-buffered saline plus 0.1% Tween 20 containing 5% fat-free milk powder for 1 h at room temperature and then incubated under the same conditions with antipeptide antiserum (1:2000) for 1 h. Immunoreactive bands were visualized by an enhanced chemiluminescence reagent system (Amersham Biosciences, Inc.). After being found to be the only one able to detect polypeptide HGB.A (see "Results"), antiserum AF22P3 (anti-FHG22-(76-95)) was used for standard immunodetection.

Enzymic Deglycosylations-- In the study of unpurified HGB deglycosylation, female hamster parotid, submaxillary, and harderian gland homogenates or saliva were digested with endoglycosidases to remove saccharides from the polypeptide chains, and the protein was then immunodetected with AF22P3. Briefly, the O-linked sugars were removed from proteins present in gland homogenates or saliva (40-80 µg of total protein/sample in adequate amounts to give similar band intensities) by incubating each of the samples in a total volume of 50 µl with 3 milliunits of neuraminidase plus 2 milliunits of O-glycosidase in 0.1 M sodium phosphate buffer, pH 6.0, for 18 h at 37 °C. Similarly, N-deglycosylation of the proteins was performed by incubating samples with 0.6 units of PNGase F and 10 mM sodium phosphate buffer, pH 6.0, in reaction volumes of 50 µl for 18 h at 37 °C. Control reactions were incubated under the same conditions but in the absence of enzymes. Samples were then analyzed by non-reducing SDS-PAGE followed by AF22P3 immunodetection.

Complete removal of (N-linked) sugars from a purified parotid gland HGB preparation was achieved by incubating 10 µg of protein with PNGase F as described before but including 1 M -mercaptoethanol in the reaction mixture. The sample was then analyzed by SDS-PAGE and visualized by staining with Coomassie Blue as explained.

Purification of HGB from Hamster Parotid Glands-- The parotid glands from two female hamsters were extracted and treated as mentioned above to obtain the parotid homogenate from which the oligomer HGB was purified. The progress of purification was monitored by immunodetection of HGB.A with AF22P3 in the preparations and fractions obtained during the procedure. The parotid homogenate was first subjected to differential precipitation in ammonium sulfate at a temperature of 0 °C. The homogenate was brought up to 50% ammonium sulfate saturation, stirred for 1 h, and centrifuged at 15,000 × g for 20 min. The precipitate was discarded, and ammonium sulfate was added to the supernatant to reach 80% saturation, stirred, and centrifuged again as described. The precipitate was collected and resuspended in 3 volumes of 10 mM sodium phosphate, pH 7.5, and this preparation was called parotid 80-50 precipitate, with a protein concentration around 10 mg/ml. A precipitate sample (600 µl) was applied to an ion exchange chromatography DEAE-Sephacel (16/20) column coupled to a Gradifrac system (Amersham Biosciences, Inc.) equilibrated with 10 mM sodium phosphate, pH 7.5, at a flow rate of 18 ml/h, and the column was washed with 1 volume of equilibration buffer. Bound proteins were eluted with 150 ml of a linear gradient from 0 to 0.5 M NaCl in the same buffer. Fractions (3 ml each) containing immunodetected HGB.A were located at a protein peak eluting between 0.3 and 0.36 M NaCl and analyzed by non-reducing and reducing SDS-PAGE; those showing contaminant proteins were discarded. Fractions of interest were pooled, dialyzed against 200 volumes of 10 mM ammonium bicarbonate, lyophilized, and finally resuspended in equilibration buffer to obtain a protein concentration around 5 mg/ml. Aliquots of this preparation were used for protein characterization by SDS-PAGE and amino acid sequencing of the subunits.

Amino Acid Sequencing-- Either 100 µg of parotid homogenate or 20 µg of purified HGB were separated by SDS-PAGE under reducing conditions, and the gels were stained with Coomassie Blue as explained. Protein bands of interest were excised and in gel digestion with trypsin performed automatically in the Progest (Genomic Solutions) as explained (47). The tryptic peptides (trim) were separated by high performance liquid chromatography (48) and sequenced in a 474 Procise peptide sequencer (Applied Biosystems).

PCR Cloning, Screening, and Isolation of cDNA Clones for HGB.B-- To obtain a partial cDNA probe corresponding to the largest HGB subunit, two reactions of degenerate oligonucleotide-primed PCR (49, 50) were prepared using a female hamster harderian gland cDNA library (37) as template. The common sense primer was a 20-mer oligonucleotide mix (5'-GAYGAYGCNATHGCNAARAC-3') including the codon choices for the internal heptapeptide DDAIAKT (see Table I), and the two antisense primers corresponded to sequences of the promoters SP6 (5'-TCAAGCTATGCATCAAGCTT-3') or T7 (5'-ACGGCCAGTGAATTGTAATA-3') flanking the multiple cloning site of the phagemid pcDNAII (Invitrogen). The amplification reactions were set at a final volume of 50 µl containing 1 µg of template DNA, 10 pmol of each primer, 200 µM each dNTP, 50 mM KCl, 15 mM MgCl₂, and 1 unit of Taq DNA polymerase (Roche Molecular Biochemicals) and performed under the following conditions: an initial denaturation step at 94 °C for 5 min before the addition of the enzyme followed by 30 cycles at 95 °C for 1 min, 48 °C for 1 min, and 72 °C for 2 min and a final elongation step at 72 °C for 5 min. The reaction corresponding to the SP6 primer produced a 350-base pair fragment that was isolated, subcloned into the plasmid pGEM-T (Promega), sequenced, and used as a probe in the screening of 5 × 10³ colonies of the female hamster harderian gland cDNA library (37). Eight positive clones were isolated and completely sequenced from both ends, revealing the existence of two highly homologous cDNAs named thereafter HGB.B1 and HGB.B2.

RNA Preparation and Northern Analysis-- RNA was extracted as previously described (3, 41) from the following male and female hamster tissues: spleen, brain, liver, small intestine, pancreas, lung, kidney, heart, thymus, adrenal gland, harderian gland, parotid gland, and submaxillary gland and also from ovary, uterus, prostate, seminal vesicle, and testes. Concentration, purity, and integrity of the RNA samples were assessed by A_260/280 measurement and agarose electrophoresis. Total RNA (20-30 µg/lane) was separated using 1% agarose-formaldehyde gels and transferred to Duralose UV^TM membranes (Stratagene) as described (41). Ethidium bromide fluorescence of the rRNAs was used to check even loading and approximately quantify samples. To determine the mRNA levels of the subunits HGB.A or HGB.B in the different tissues, Northern blots were hybridized as described with [-³²P]dCTP-labeled cDNA probes in the presence of 40% formamide at 42 °C using, respectively, the HGB.A-FHG22 cDNA (3) or a combination containing equimolar amounts of HGB.B1 and HGB.B2 cDNAs. Specific oligonucleotide probes with 20% mismatch to the opposite sequence were labeled with [-³²P]ATP and polynucleotide kinase (Roche Molecular Biochemicals) and used to detect either HGB.B1 (5'-GTAAATGGCAAACTCCATCA-3') or HGB.B2 (5'-TTGTAAACCGCATACACCAT-3') mRNAs by hybridization in the presence of 0.3 M NaCl, 0.3 M sodium citrate, 5× Denhardt's solution, 50 mM phosphate buffer, pH 6.6, 0.1% SDS, and 0.1 mg/ml yeast tRNA at 52 °C (4 °C below the T_m) as described by Sambrook et al. (51). The specificity and detection range of the oligonucleotides were determined using identical conditions with Southern blots having 0.01, 0.1, 1, and 10 ng of each HGB.B1 and HGB.B2 cDNAs run in adjacent lanes. Blots were washed thoroughly to eliminate the radioactivity when hybridized successively to more than one probe.

Sequence Comparisons-- The nucleotide and amino acid sequences of HGB.B1, HGB.B2, and HGB.A were compared among themselves and also with the cDNA-deduced polypeptide sequences of available SCGB family members (11, 13) using the ClustalX application (52). A phylogenetic tree with subfamilies generated by homology was obtained, and the multiple alignments of the subfamilies including the HGB subunits are represented.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

HGB.A Is the Mature Polypeptide Encoded by the FHG22 mRNA-- Northern analysis of the FHG22 mRNA (0.6 kilobases) demonstrated a tissue-specific and sex differential expression in Syrian hamster (3). The mRNA expression pattern was properly reassessed here (Fig. 1A) and used as reference in the immunodetection of the expected polypeptidic product. For this purpose, three antipeptide antisera corresponding to amino, central, and carboxyl parts of the mature polypeptide sequence deduced from the cDNA (3) were prepared by rabbit immunization and used for Western analysis of hamster samples after SDS-PAGE in reducing conditions. Under these conditions only the antiserum AF22P3 (raised against the carboxyl-terminal peptide FHG22-(76-95) was found to detect a unique polypeptidic band with an apparent size of 5.6 kDa (Fig. 1, B and C). The band is found in harderian, parotid, and submaxillary homogenates and also in saliva; the highest levels appear in male and female parotid glands, whereas in submaxillary glands and in saliva, the levels are higher in females and absent in male harderian glands (Fig. 1B). This pattern of expression is the same as that of FHG22 mRNA such that a direct correlation between them is observed. The correlation is further supported by the fact that the band is not observed when using AF22P3 antiserum for the analysis of FHG22 mRNA negative tissues such as male and female thymus and is also undetected in other biological fluids such as female serum or vaginal discharge (data not shown).

View larger version (40K): [in this window] [in a new window]   Fig. 1.   Identification of polypeptide HGB.A as the product of FHG22 mRNA. Hamster tissues and saliva were used for detection of mRNA and/or protein and for amino-terminal sequencing. A, 30 µg of total RNA obtained from female (f) and male (m) harderian, parotid, and submaxillary glands were subjected to Northern blot analysis and hybridized with a FHG22 cDNA probe. B, tissue homogenates or saliva as indicated (50 µg of protein/lane) were treated with 1.4 M -mercaptoethanol for 30 min and subjected to 15% SDS-PAGE followed by immunodetection using AF22P3 antiserum. C, two female parotid homogenate samples (100 µg each) were subjected to SDS-PAGE as in panel B and used for immunodetection with AF22P3 (lane 1) or stained with Coomassie Brilliant Blue (lane 2). The band containing the polypeptide identified as HGB.A was trimmed from the gel and subjected to automated Edman degradation. The amino-terminal sequence of the polypeptide and the open reading frame found in FHG22 cDNA are compared in the figure.

For obvious reasons explained below, the 5.6-kDa polypeptide will be called HGB.A from now on. To demonstrate that it is indeed encoded by the FHG22 mRNA, the amino-terminal sequence of HGB.A was determined as follows. Two female parotid homogenate samples (100 µg of protein each) were separated by reducing SDS-PAGE, and the gel was divided; one sample was stained with Coomassie, and the other was transferred to a nitrocellulose membrane and used for immunodetection (Fig. 1C). The comparison of the Western signal with the protein profile demonstrates that the AF22P3-detected HGB.A band (lane 1) comigrates with a polypeptide highly expressed in parotid gland (lane 2). Because of its apparent lack of contamination by other polypeptides, the band was excised from the gel and used for sequencing; the only amino-terminal sequence obtained was ANVCPAVLSVS(K), thus confirming the identity and purity of the band. As reflected in Fig. 1C, it exactly matches residues 22-33 of the ORF found in the FHG22 cDNA, thereby demonstrating that the predicted signal peptidase cleavage site was correct (3, 53) and identifying the HGB.A polypeptide as the mature product of FHG22 mRNA. A doubt as to its positive identification could be raised by the clear difference observed between the cDNA-deduced size of the mature product (M_r 8,196) and that measured by SDS-PAGE (M_r 5,600). Using Tris-Tricine buffers, reported to be more appropriate for discrimination of small polypeptides (54, 55), the apparent size increases up to M_r 7,000 (data not shown).

HGB.A Is a Subunit of the Disulfide-bound Heterodimer HGB-- All members of the SCGB family have been reported as or supposed to be subunits of disulfide-bound dimers (10-13). When hamster tissue homogenates or saliva were analyzed in non-reducing conditions, only a 37-39-kDa band (Fig. 2A) showing identical expression profile (data not shown) was observed instead of HGB.A. This indicates that in native biological samples HGB.A is found only as part of oligomeric protein(s), the nature of which was investigated before proceeding to its purification. When a parotid homogenate sample is electrophoresed in semi-reducing conditions and analyzed by immunodetection, the 39-kDa and HGB.A bands are simultaneously observed (Fig. 2B, lane 1). Both structures are specifically recognized by AF22P3 since coincubation with the antigenic peptide FHG22-(76-95) (lane 2), but not with FHG22 55-66 (lane 3), impedes immunodetection of the bands. These data suggest that the 39-kDa band is a disulfide-bound oligomer that renders free HGB.A upon reduction; because no additional bands are detected in semi-reducing conditions, HGB.A is likely bound to only one counterpart, and the oligomer is a heterodimer. However, some related proteins are formed by two non-covalently bound heterodimers (8, 14-16), and for this reason we determined the size of the oligomer in native conditions by gel filtration chromatography. A female parotid homogenate sample was eluted through a Sephacryl S-100 column, and the presence of HGB.A was immunodetected in the fractions and compared with the elution profile of molecular weight standards. Using this method, HGB.A was only detected in fractions corresponding to a protein peak with an apparent M_r of 34,000 (data not shown), which demonstrates the lack of non-covalently bound heterotetramers and, hence, that the protein band detected by SDS-PAGE accounts for the complete oligomeric structure. However, some experimental observations indicated that the oligomer should be studied in the three expressing tissues. In fact, a more accurate determination of the band size in female homogenates after long-run SDS-PAGE in non-reducing conditions (Fig. 2A) permits visualization of the differences between the apparent sizes of the oligomer from submaxillary (M_r 37,000) and those from parotid and harderian glands (M_r 39,000), thereby suggesting that HGB.A can be bound to different counterparts. Because there is no defined physiological function, the name heteroglobin was given to these oligomeric proteins having the polypeptide HGB.A as subunit because of their heterodimeric structure, heterogeneity in size and biochemical composition, and heterotypic tissue expression, as shown in this work.

View larger version (66K): [in this window] [in a new window]   Fig. 2.   Immunodetection of oligomeric HGB and monomeric HGB.A. Female hamster tissue homogenates (50 µg of protein per sample) were subjected to non-reducing or semi-reducing SDS-PAGE and used for immunodetection with AF22P3 antiserum. A, homogenates from female parotid, submaxillary, and harderian glands were subjected to long-run 10% SDS-PAGE in non-reducing conditions and used for immunodetection with AF22P3. B, three parotid homogenate samples were treated with 0.14 M -mercaptoethanol 30 min, subjected to 15% SDS-PAGE, and used for immunodetection with AF22P3 antiserum incubated in the absence (lane 1) or presence of 1.25 µg/ml of peptide FHG22-(76-95) (lane 2) or FHG22-(55-66) (lane 3).

Deglycosylation of HGB-- The observation that parotid HGB presents different sizes when measured by SDS-PAGE or chromatography could be due to a possible glycoproteinic nature (55, 56). To detect the presence of saccharide chains in the molecule, female gland homogenates or saliva were digested either with neuraminidase plus O-glycosidase or with PNGase F and subjected to non-reducing SDS-PAGE followed by immunodetection with AF22P3 (Fig. 3). Removal of O-linked sugars does not affect HGB molecules, since the band pattern from tissues and saliva (Fig. 3B) is equivalent to that found in undigested control incubations (Fig. 3A). Rather, treatment with PNGase F alters migration of HGB molecules (Fig. 3C); in parotid glands, harderian glands, and saliva, the apparent size is reduced in 2 steps (from 39,000 to 33,000 and 25,000), suggesting that two N-linked oligosaccharide branches are being removed, whereas in submaxillary glands only one size reduction is clearly observed. Although this is positive proof of the presence of Asn-bound oligosaccharide chains in the HGB molecules, this procedure cannot be used for complete deglycosylation analysis due to the fact that full PNGase F action is only achieved in reducing conditions not compatible with detection using AF22P3. Because HGB.A cannot have a carbohydrate moiety because of its size, absence of consensus sequences for N-glycosylation, and the fact that no change in the band was observed after treatment with PNGase F (see Fig. 4), the oligosaccharide chains must be N-linked to the other subunit of the HGB molecule.

View larger version (39K): [in this window] [in a new window]   Fig. 3.   Partial deglycosylation of HGB from tissue homogenates and saliva. Female hamster tissue homogenates or saliva in adequate amounts to produce similar band intensities in Western analysis (40-80 µg of protein/lane) were incubated in the absence of enzymes (panel A), with 3 milliunits of neuraminidase (Neur.) plus 2 milliunits of O-glycosidase (O-gly.; panel B) or with 0.6 units of peptide N-glycosidase F (panel C) for 18 h at 37 °C. Samples were then subjected to 15% SDS-PAGE in non-reducing conditions followed by immunodetection using AF22P3 antiserum.

View larger version (54K): [in this window] [in a new window]   Fig. 4.   Purification of HGB from hamster parotid glands. Female parotid homogenate was used as a source for purification of HGB. Advance of the process was monitored by immunodetection with AF22P3 antiserum after each step. Samples obtained along the purification process were analyzed by 15% SDS-PAGE in non-reducing (lanes 1-4) or reducing conditions (lanes 5 and 6) and stained with Coomassie Blue. Lane 1, parotid homogenate (70 µg); lane 2, ammonium sulfate precipitate (40 µg); lane 3, ion exchange-purified HGB (10 µg); lane 4, markers; lane 5, purified HGB (10 µg) treated with 1.4 M -mercaptoethanol for 30 min; lane 6, purified HGB (10 µg) treated with 0.6 units of peptide N-glycosidase F and 1 M -mercaptoethanol for 18 h.

Purification and Characterization of HGB-- Parotid gland (from females) was used as a source for purification of the heterodimer due to the fact that it shows the highest levels of HGB.A (see Fig. 1 and 2) and convenient protein composition. A simple protocol followed by immunodetection of HGB.A after each step permits the purification of HGB from parotid homogenate in which it is a major protein, as shown by electrophoretic analysis (Fig. 4). Briefly, parotid homogenate (lane 1) was subjected to differential precipitation with ammonium sulfate; the precipitate (lane 2) was chromatographed through an ion-exchange column, and purified HGB was obtained from selected fractions (lane 3). This preparation was used to determine the size and subunit composition of the oligomer by SDS-PAGE. As expected, it migrates as a 39-kDa band in non-reducing conditions (lane 3), but after thiol-reducing treatment (lane 5), two bands are observed; the small one (5.6-kDa) was named subunit HGB.A as previously mentioned, and the large one (33.5 kDa) was named subunit HGB.B. Complete digestion of HGB with PNGase F (in reducing conditions) only affects HGB.B, whose apparent size decreases from M_r 33,500 to 28,000 to 19,000 (data not shown) and finally to 9,000 (lane 6). Thus, the carbohydrate part accounts for most of the HGB.B molecule, which is likely to have three N-linked oligosaccharide chains. Finally, the purity of the protein preparation was also analyzed by isoelectric focusing; oligomeric HGB migrates as a unique band with a very acidic pI around 2.8, in agreement with the expectable acidic nature of the carbohydrate moiety (data not shown).

Partial Sequencing of the Subunits HGB.A and HGB.B-- Purified parotid gland HGB was subjected to reducing SDS-PAGE, and the bands corresponding to the subunits HGB.A and HGB.B were processed for protein sequencing as described (47, 48). As expected, amino-terminal sequencing of HGB.A elicited an identical sequence to that found in the band from parotid homogenate (see Fig. 1C). The results of sequencing the amino terminus and tryptic peptides of the HGB.B band are shown in Table I, such that a 33 residue-long stretch including the former could be reconstituted. The lack of Met in the first position indicates that, as in HGB.A, the signal peptide has been removed, and the sequence corresponds to a mature polypeptide. Two consensus N-glycosylation sites are observed at positions 19 and 35, in which the Asn residues are probably true glycosylation sites as there is a gap in the sequencing signal (57). All the tryptic peptides can be located in the reconstituted amino-terminal sequence and/or identified after Lys residues in the two cDNA-deduced sequences shown in Fig. 5, including the artifactual sequence of trim 17 which is composed of amino acids 91-94 followed by 77-82 (Table I). The positive identification of two peptides differing in one amino acid residue (trim 30 YTTLPYIR and trim 33 YTFLPYIR) corresponding to the sequences present in HGB.B2 and HGB.B1 (Fig. 5) demonstrates the presence of both polypeptides in parotid glands, which is also supported by the detection of residues from HGB.B2 (Pro-39) and HGB.B1 (Gln-48) in the sequence of trim 21. Finally, the codon combinations of the amino-terminal sequence of HGB.B were analyzed and utilized to obtain a DNA probe by mixed oligonucleotide-primed amplification of cDNA. The probe was successfully obtained using a mix of 384 oligonucleotide sequences (20-mers) that included all the possible codons for the sequence DDAIAKT except the most 3' position (see next paragraph).

View this table: [in this window] [in a new window]   Table I Amino acid sequencing of the HGB.B subunit from parotid HGB Purified HGB (20 µg) was treated with 1.4 M -mercaptoethanol 30 min, and subjected to 15% SDS-PAGE, and the gel was stained with Coomassie Blue. The band containing the 34.5-kDa polypeptide (HGB.B) was excised from the gel and used for amino-terminal and tryptic peptide sequencing. "Position in ORF refers" to the first and last residues in the open reading frames reported in Fig. 5. "N-term" indicates an extended amino terminal sequence including trim 21. Underlined residues indicate two differential amino acids positively identified in trim 30 and trim 33 and the sequence used to design a degenerate oligonucleotide mix.

View larger version (57K): [in this window] [in a new window]   Fig. 5.   Nucleotide sequences and corresponding amino acid translations of HGB.B1 and HGB.B2. Nucleotide and amino acid (in boldface) numbers are counted by reference to the first position of initiating ATG codons as +1. Lowercase letters denote nucleotide differences, and italic uppercase letters denote amino acid differences. Underlined nucleotides show the polyadenylation signal in the 3'-untranslated region of the two cDNAs and two differential restriction sites mentioned under "Results." Underlined amino acids indicate N-glycosylation motifs. Signal peptides are boxed, and relevant amino acids are circled. The HGB.B1 and HGB.B2 cDNA sequences have been entered into the EMBL/GenBankTM/DDBJ data bases and are found under GenBankTM accession numbers AJ252138 and AJ252139, respectively.

Isolation and Characterization of cDNAs for HGB.B1 and HGB.B2-- A female hamster harderian gland cDNA library (37) was used as the template in degenerate oligonucleotide-primed PCR amplifications (49, 50), with the oligonucleotide mix defined by 5'-GAYGAYGCNATHGCNAARAC-3' as sense primer and two antisense primers specific for the cloning vector. A 350-base pair-long DNA fragment was successfully amplified, cloned, sequenced, and found to harbor an ORF starting with the sequence DDAIAKT and agreeing with the sequences described in Table I. It was then used as a probe to screen the female harderian gland cDNA library such that two highly homologous cDNAs could be identified (Fig. 5). Eight positive clones were isolated and sequenced during the process, seven of which are identical (HGB.B1), whereas the other (HGB.B2) was very similar (96.3% homology) and identical to the (partial) sequence of the cloned PCR product. A 477-nucleotide-long sequence containing a polyadenylation signal and followed by a poly(A) tail was obtained from the HGB.B1 clones, whereas the HGB.B2 sequence obtained was seven and eight nucleotides shorter at the 5' and 3' ends, respectively (Fig. 5). Parallel ORFs are observed in both cDNAs, starting at a Met with consensus sequence for translational initiation (58) and coding for 94-amino acid-long sequences showing 87.6% identity and 93.6% similarity (52). Comparison with the HGB.B amino terminus (Table I) revealed identical signal peptides with standard cleavage sites (53) between Cys-16 and Arg-17 in both sequences (Fig. 5). Remarkably, the untranslated and signal peptide nucleotide sequences are also identical, such that all the 17-base differences found are restricted to the 78 codons of the mature polypeptides, producing 12 amino acid substitutions (7 conservatives) between mature HGB.B1 and HGB.B2 (Fig. 5). Despite these changes, three consensus sites for N-glycosylation are conserved at residues 19, 35, and 72, in agreement with the deglycosylation data from Figs. 3 and 4 and the gaps detected in the sequencing process. Also, both polypeptide sequences show three residues conserved in the SCGB family, Cys-23, Lys-64, and Cys-91, as well as Cys-66, conserved in all heterodimeric members of the family (11-13). According to their cDNA sequences, the calculated M_r of non-glycosylated HGB.B1 and HGB.B2 polypeptides are, respectively, 10,883 and 10,821 before and 9,048 and 8,986 after the action of the signal protease, in accordance with the apparent size of the deglycosylated HGB.B band (see Fig. 4).

Parallel Expression of HGB.B and HGB.A mRNAs-- To investigate whether the expression of both HGB subunits is transcriptionally coordinated, we determined the mRNA levels for HGB.B and HGB.A in a broad collection of RNAs from male and female hamster tissues by successively probing the same blots with equimolar HGB.B1/B2 or HGB.A cDNA probes as explained under "Experimental Procedures." Homology between HGB.A and either HGB.B1 or HGB.B2 cDNA sequences (52) is around 47%, such that no cross-hybridization with mRNAs was expected in the conditions used. No expression for any HGB mRNA was found in male and female adrenal glands, brain, heart, kidney, liver, pancreas, skeletal muscle, small intestine, spleen, and thymus or in ovary, seminal vesicle, and testes (data not shown). The result of hybridizing the HGB.B and HGB.A probes to RNA from some tissues of interest is shown in Fig. 6. Not surprisingly, the mRNAs showed almost indistinguishable sizes around 0.6 kilobases and very similar expression patterns: high levels in parotid glands from both sexes, lower levels with sexual differences (more in females) in submaxillary glands, and female-specific expression in harderian glands; no expression was observed in male harderian gland, prostate, uterus, or lung from any sex. Indeed, the patterns clearly concur with the protein distribution previously described (see Fig. 1 and 2), and it is interesting to highlight that no tissue with independent expression of HGB.A or HGB.B mRNAs was detected.

View larger version (80K): [in this window] [in a new window]   Fig. 6.   Parallel expression of HGB.B and HGB.A mRNAs in hamster tissues. Total RNA was extracted from male and female hamster tissues, and 30 µg of each preparation were electrophoresed in agarose-formaldehyde gels and transferred to nitrocellulose membranes. Membranes were used for detection of HGB.B mRNAs, washed thoroughly, and rehybridized for HGB.A mRNA detection. The HGB.B probe consisted of an equimolar mixture of HGB.B1 and HGB.B2 cDNAs, and the HGB.A probe was FHG22 cDNA (3). In addition to the tissues shown in the figure, no expression was detected in adrenal glands, brain, heart, kidney, liver, pancreas, skeletal muscle, small intestine, spleen, and thymus from any gender nor in ovary, seminal vesicle, and testes. The bottom panel shows ethidium bromide staining of ribosomal RNA corresponding to the indicated tissues.

Differential Expression of HGB.B1 and HGB.B2 mRNAs-- Recurrent observation of such a specific expression pattern prompted us to determine the particular contribution of HGB.B1 and HGB.B2 mRNAs to the HGB.B pool. New RNA samples from parotid, submaxillary, and harderian glands were prepared, blotted, and successively hybridized to three probes, one able to detect both mRNAs (equimolar combination of HGB.B1 and HGB.B2 cDNAs) and two differentials able to detect each mRNA (specific oligonucleotides with 20% mismatch to each other's mRNA); the results are shown in Fig. 7. When using the combined cDNA probe, the pattern already described for total HGB.B (see Fig. 6) was essentially repeated as expected. However, hybridizations of the RNA blot to each of the specific oligonucleotides demonstrate that HGB.B1 is expressed in parotid and submaxillary glands, whereas HGB.B2 is clearly present in parotid but not in submaxillary glands. To avoid misinterpretations, the ability of the specific probes to detect low amounts (10,000 target sequences in 0.01 ng of cDNA; not shown) of proper nucleic acid without showing cross-hybridization to higher amounts of the opposite is shown in the control Southern analysis in Fig. 7. Unfortunately, the very low level of HGB.B mRNAs found in these female harderian glands impeded detection using these oligonucleotide probes; low HGB expression not related to estral cycle, age, or main environmental factors has occasionally been observed in female harderian glands.² However, using a reverse transcription-PCR-based procedure that takes advantage of two differential restriction sites (NsiI at +71 in HGB.B1 and ScaI at +172 in HGB.B2; see Fig. 5), we were able to demonstrate the presence of both HGB.B1 and HGB.B2 mRNAs in these female harderian glands. The same analysis confirmed that in submaxillary glands from either sex HGB.B1, but not HGB.B2 mRNA, was found, whereas in parotid glands, HGB.B1 and HGB.B2 existed as expected (data not shown).

View larger version (67K): [in this window] [in a new window]   Fig. 7.   Differential expression of HGB.B1 and HGB.B2 mRNAs. Total RNA was extracted from the indicated tissues of female (f) and male (m) hamster, and about 30 µg were subjected to Northern analysis. The membrane was successively hybridized to the probes indicated in the figure and stripped of the radioactivity each time. The common probe contained equimolar amounts of HGB.B1 and HGB.B2 cDNAs, and the oligonucleotide probes had sequences specific for HGB.B1 or HGB.B2, showing 20% mismatch to the opposite mRNA sequence. The ability of each nucleotide to detect the indicated amounts of proper cDNA without showing cross-hybridization to the other sequence is illustrated in the Southern blots from the right panel. The bottom panel shows the ethidium bromide staining of ribosomal RNA corresponding to the indicated tissues.

Both HGB Subunits Belong to the SCGB Family-- Previous descriptions of the UGB-SCGB family include as members HGB.A (under the name of FHG22) and two other cDNA-deduced polypeptide sequences submitted by us to GenBank^TM (accession numbers AJ252138 and AJ252139) with the name "heteroglobin" (10, 12, 13). No evidence was made available for the existence of the corresponding polypeptides, but in this article they have been described as the subunits of HGB. Family members have been found to be around 90-95 amino acids long, including a signal peptide and two conserved Cys, located at both ends of the mature polypeptides according to our data for HGB.A, HGB.B1, and HGB.B2. We developed a family tree for 32 members of the SCGB family (52) in which five subfamilies were segregated by homology (data not shown), in agreement with other authors (10, 13). Because a definitive nomenclature has yet to be established, subfamilies including HGB.A and HGB.B have been named after them, and their multiple sequence alignments are shown in Fig. 8. HGB.A subfamily also includes lipophilins type A and B from man and rabbit (8, 9, 11), human lymphoglobin (13), and the C1 and C2 subunits of rat prostatein (59). The HGB.B subfamily includes the two isoforms described in this paper, the two sequences reported for the prostatein C3 subunit (60, 61), human mammaglobin (5) and lipophilin C (or mammaglobin B or lacryglobin) (6, 8, 62), and rabbit lipophilins type C (11). The three residues conserved in all the members of the SCGB family (Cys-23, Lys-64, and Cys-91; see Fig. 5 numbering) are shown in bold in Fig. 8. Cys-23 and Cys-91 are thought to be responsible for the formation of interchain disulfide bridges in all SCGBs, and Lys-64 is positioned in the calcium binding site (13, 30), whereas Cys-66 is only conserved in heterodimeric members and has been proposed to be involved in the formation of an additional disulfide bridge (8, 12). Although a common structure with four -helices seems to be maintained in the whole family (12, 13), different residues are almost or absolutely conserved in each subfamily as illustrated (Fig. 8). Finally, it is worth mentioning that all HGB.B subfamily members show at least one N-glycosylation site, but only HGB.B1 and HGB.B2 have three.

View larger version (41K): [in this window] [in a new window]   Fig. 8.   Multiple sequence alignments of SCGB subfamilies including the two HGB subunits. An updated SCGB family phylogenetic tree containing five subfamilies was generated using cDNA-deduced amino acid sequences and Clustal X. Alignments corresponding to the two subfamilies including HGB.A or HGB.B polypeptides are shown. Each sequence is named with the species initials (Hs, Homo sapiens; Ma, M. auratus; Oc, Oryctolagus cuniculus; Rn, Rattus norvegicus) and the abbreviation for the common name found in the literature. Amino acid sequences derive from the cDNAs corresponding to the following GenBankTM accession numbers: a, AJ224171; b, AF308616; c, AJ224172; e, AF308614; f, AF308615; g, Z66540; h, J00774; i, J00776;  j, AJ224173; k, U33147; l, AF308617; m, AF308618; n, AF308620; o, AF308619; p, V01263; q, V01263; r, AJ252138; s, AJ252139. Sequence d, corresponding to lymphoglobin (Hs YGB), is reported in Ni et al. (13). Numbering refers to the complete open reading frames. Amino acids totally conserved in the family are shown in bold. Conservation of amino acids in each subfamily is indicated as follows: asterisk, residues totally conserved; dash, residues partially conserved. N-Glycosylation sites or motifs found in HGB.B subfamily members are underlined.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

This paper describes the identification and characterization of two isoforms of hamster HGB, a heterodimeric protein whose disulfide-bound subunits (HGB.A and HGB.B) belong to the expanding SCGB-UGB family (1-4, 10). The common small subunit HGB.A was first detected using antipeptide antiserum and later identified by sequencing as the mature polypeptide encoded by the tissue- and sex-specific hamster FHG22 mRNA (3). Indeed, immunodetection of HGB.A permitted purification of heterodimer from parotid gland, which in turn led to the cloning of two cDNAs coding for HGB.B1 and HGB.B2, the two isoforms of the N-glycosylated large subunit, and the study of their differential expression. With the lack of clues as to its function(s), the protein was named heteroglobin, thus reflecting heterogeneity because (i) it is a heterodimer, (ii) it is a glycoprotein, (iii) it has at least two isoforms, and (iv) subunits and isoforms show tissue and sex differences in expression. Additionally, the suffix "globin" has been chosen for the SCGB family, implying dimerization behavior to produce a conserved eight-helix bundle structure surrounding a hydrophobic pocket (10, 12).

The logical procedure after cloning FHG22 was to identify the encoded polypeptidic product; antiserum raised against the carboxyl terminus was able to immunodetect a 5.6-kDa band showing the same tissue and sex distribution as FHG22 mRNA and also present in saliva (Fig. 1). The comparison of amino-terminal and deduced sequences (Fig. 1C) demonstrates that it is in fact the encoded polypeptide, named HGB.A, after having the signal peptide removed according to the cleavage site previously proposed (3). The difference between apparent (M_r 5,600) and calculated (M_r 8,196) sizes is attributable to irregular migration of HGB.A, a technical trouble also reported for other polypeptides of the SCGB family (4, 8, 14, 16). Indeed, integrity of the polypeptide is supported by identification of the amino and carboxyl termini by sequencing and immunodetection, respectively. The presence of HGB in parotid and submaxillary glands explains its detection in saliva (Fig. 2 and 3) to which the protein must be secreted; indeed, it has been reported that hamster harderian gland secretion may contribute to salivary composition (39). The higher levels detected in female versus male submaxillary and harderian glands concur with the estradiol activation of HGB.A mRNA expression reported for the latter (20), but such a difference is not observed in saliva, in which variations in HGB levels have been studied in males and females but could not be related to hormonal or environmental factors (data not shown).

Despite the size difference observed in submaxillary glands, it is clear that HGB from any source is N-glycosylated at least twice, as shown in Fig. 3. Experiments performed with purified HGB demonstrate that all the sugar chains are bound to the large subunit HGB.B (M_r 33,500) as expected, whose core polypeptide showed an apparent M_r of 9,000 after complete PNGase F treatment (Fig. 4), in accordance with the sizes calculated for mature HGB.B1 and HGB.B2. Amino acid sequencing of the HGB.B band also supported the presence of at least two Asn-bound sugar chains (Table I). Very likely, three oligosaccharide chains are N-linked to HGB.B, in agreement with observed deglycosylation bands (data not shown), and with the three N-glycosylation sites conserved in HGB.B1 and HGB.B2 (Fig. 5). However, the partial and average size differences observed during deglycosylation of HGB (Fig. 3) and HGB.B were higher than the expected size of most N-linked oligosaccharides (M_r 5,500-10,000 versus 3,000-4,000), perhaps due to alterations in mobility produced by charge shifts after sugar branches were removed from a small polypeptide (55, 56). Anomalous mobility could also account for the difference between apparent sizes of HGB from the parotid gland, measured by SDS-PAGE (M_r 39,000) or by gel filtration chromatography (M_r 34,000), since it has been described that highly glycosylated proteins show an irregular behavior in SDS-PAGE gels (55, 56). The fact that the protein is easily purified from a differential ammonium sulfate precipitate (Fig. 4) may also be due to the presence of sialic acids and sulfate esters that make some glycoproteins bind well to ion-exchange columns (63).

Besides the simultaneous expression of HGB.A, HGB.B1, and HGB.B2 mRNAs in parotid glands (Figs. 6 and 7), proofs of the existence of HGB.A-HGB.B1 and HGB.A-HGB.B2 isoforms of HGB rely on protein sequencing of HGB.A and HGB.B, the latter showing the presence of tryptic peptides containing residues of HGB.B1 and HGB.B2 (Table I). This raises the hypothesis that the whole quaternary structure could be formed by two non-covalently associated heterodimers, A-B1 and A-B2, like C1-C3 and C2-C3 in prostatein (14). This possibility must be ruled out since the size of the oligomer from parotid glands measured by gel filtration chromatography in native conditions (M_r 34,000) is equivalent to the size of the heterodimer measured by SDS-PAGE (M_r 39,000). Furthermore, such a tetrameric structure would imply similar tissue levels of HGB.B1 and HGB.B2, which is obviously not possible in submaxillary glands due to the lack of HGB.B2 mRNA (Fig. 7). Isolation of seven HGB.B1 and one HGB.B2 cDNA clones from the female harderian gland library supports the reverse transcription-PCR data showing that both mRNAs must be expressed but suggests that HGB.B1 is so to a higher extent, which is also in disagreement with the existence of an A-B1:A-B2 oligomer in harderian glands.

The distribution of nucleotide differences between the HGB.B1 and HGB.B2 cDNA sequences indicates that they correspond to similar isoforms encoded by genes from different loci instead of alternatively spliced mRNAs or two expressed alleles. Thus, although the 5'- and 3'-untranslated and the signal peptide sequences are identical (Fig. 5), it is remarkable to note that inside the 234 nucleotides encoding the mature polypeptides there are 17 differences affecting two exons (data not shown). The fact that the genes encoding HGB.B1 and HGB.B2 are differentially expressed at the transcriptional level (Fig. 7) precludes the possibility that they might be alleles from a unique locus. Also, Southern analysis of hamster genomic DNA using different restriction endonucleases shows that an HGB.B cDNA probe hybridizes to two or more fragments, suggesting the existence of different loci.³ Similarly, two rat C3 isoforms differing in six amino acid residues are encoded by genes from distinct loci (60, 61). Comparison of the cDNA-deduced sequences of the HGB subunits with those included in the SCGB family (11, 13) permits us to demonstrate that HGB.A belongs to subfamily D, whereas HGB.B1 and HGB.B2 belong to subfamily C according to the nomenclature of Ni et al. (13). Several authors report the existence of heterodimers formed by members of subfamily D disulfide-bound to N-glycosylated counterparts from subfamily C, such as subunits C1-C3 and C2-C3 from rat prostatein (14), lipophilins A and C in human tears (8), and mammaglobin and BU101 (or lipophilin B) in human mammary gland (64). According to these observations, it has been hypothesized that a structural association exists between members of those subfamilies, and it has also been proposed that one particular SCGB polypeptide can form different heterodimers depending on the available counterparts expressed in the tissue (11, 13). For instance, polypeptide C3 forms heterodimers with C1 and C2 in prostatein but has been found to be part of a different protein in rat lacrimal and submaxillary glands (65-66). Studies on the tissue-specific expression patterns of several members of the two subfamilies support both the necessary co-expression and the possibility of forming different heterodimers (9, 13, 62, 64); like the HGB subunits, most of these SCGBs are expressed in orbital and/or salivary glands. Indeed, the HGB expression data presented in this work support the hypothesis of the necessary association between members of subfamily D and subfamily C, since HGB.A and HGB.B are not independently expressed, whereas HGB.B1 and HGB.B2 can be. The fact that A-B2 heterodimers cannot be formed in submaxillary glands, whereas A-B1 and A-B2 are detected in parotid glands, also supports the promiscuity in association according to the expressed SCGBs.

The lack of HGB.B2 mRNA in submaxillary glands suggests that both HGB isoforms might play a different physiological role despite the fact that both can be supplied to the saliva through the parotid glands. Because no clear function has been established for heterodimeric SCGBs beyond lipid binding, we are not tempted to hypothesize any physiological role. The absence of clear orthologs among them (unlike UGB), mostly exocrine patterns of expression and promiscuity in heterodimeric associations, could be related to a species-specific role(s) and/or a widely maintained structural feature such as lipid affinity. The fact that HGB shows sexual differences in expression might support both possibilities, for example by binding pheromones.

	ACKNOWLEDGEMENT

Special thanks to Dr. Antonio Nieto for permanent encouragement and useful comments.

	FOOTNOTES

^* This work was supported in part by Dirección General de Investigación Científica Técnica (DGICYT) Grant PB95-1044 from the Spanish Government.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBank^TM/EBI Data Bank with accession number(s) AJ252138 and AJ252139.

^§ Recipient of a Fundación para el Fomento en Asturias de la Investigación Científica Aplicada y la Tecnología (FICYT) fellowship from the Principado de Asturias Government.

^** To whom correspondence and reprint requests should be addressed. Tel.: 34-8-5104212; Fax: 34-8-5103157; E-mail: pedomin@correo.uniovi.es.

Published, JBC Papers in Press, October 29, 2001, DOI 10.1074/jbc.M106678200

² J. Alvarez and P. Domínguez, unpublished observations.

³ J. Viñas, J. Alvarez, and P. Domínguez, manuscript in preparation.

	ABBREVIATIONS

The abbreviations used are: SCGB, secretoglobin; UGB, uteroglobin; HGB, heteroglobin; HGB.A, HGB.B, HGB.B1, HGB.B2, subunits heteroglobin A, B, B1, and B2; ORF, open reading frame; PNGase F, peptide N-glycosidase F; Tricine, N-[2-hydroxy-1,1-bis(hydroxymethyl) ethyl]glycine; trim, tryptic in matrix digest.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES